7

Women as Labor Control

In discussing employers’ motivations for hiring women, women are customarily referred to as a “cheap docile labor force.” That women work for low wages is well known. However, whether women inhibit labor conflict is a complex and more subtle question. It is well documented that women are less likely to join labor unions (Antos et al. 1980; P. Hunt 1981). There are suggestions that women are less likely to strike. However, the interpretation of these differentials is controversial.

There are three ways to conceptualize gender differences in labor militancy. The effect of gender could be direct, indirect, or spurious. A direct effect would be one due to differential behavior by women themselves. Women for whatever reason would be inherently less militant in any conceivable work setting. An indirect effect is one created by the presence of women, not necessarily involving any distinctive action on their part. If male co-workers were to change their behavior on account of women, this would be an indirect effect. A spurious effect has no basis in gender per se. Such differences are created by women’s being disproportionately distributed into occupations or work settings that induce passivity. Any worker recruited into such positions would replicate the behavior observed in women. The fundamental question of the interrelation of gender to class conflict is the decomposition of the observed gender gap in militancy into these three components. One could then empirically determine the comparative importance of female behavior, male non-cooperation, and occupational distribution.

Most of the direct effects that have been suggested explain women’s non-participation in unions by invoking conflicts arising from traditional sex roles. Some authors have viewed women as marginal and temporary labor force paticipants. Women could escape the deprivations of the workplace by retreating into full-time family life. Furthermore, women’s short anticipated careers lowered their interest in the long-term prospects of their occupation (Bliss and Andrews 1911; Barkin 1961; Blackburn 1967; Blum 1971; Moore and Newman 1975; Hirsch 1980). Alternatively, in some industries, female employment has been restricted to juvenile labor. The young age distribution of female employees is sometimes held to produce this “short-horizon” passivity (Tilly 1981). Traditional sex roles are also seen as inhibiting the union participation of married women as well. Union activities make time demands that are often inconsistent with domestic obligations or child care (Wertheimer and Nelson 1975; P. Hunt 1981).

Changing sex roles have reduced the saliency of these arguments. Rising labor force participation has decreased the supply of passive future housewives. Housework may be playing a smaller role in determining women’s use of time. Yet the present liberation of women has not made debates about direct effects moot. Most advocates of strong direct effects would argue that changing women’s societal roles is a necessary pre-condition to militancy. Raising gender consciousness is an important pre-condition for raising class consciousness. An opponent of the direct effect thesis would claim that changing sex roles has little impact on creating female labor militancy. The only way to increase female militancy would be for them to move into jobs with more intrinsic bargaining power.

Indirect effects stem from male workers’ unwillingness to cooperate with female workers to obtain common benefits. Established male unions have in general been quite slow about organizing female workers. Those attempts that have occurred have often foundered on the simultaneous campaigns by these workers to win sexist demands such as restrictions on female labor. The result has been either the neglect or the alienation of otherwise willing female unionists (Wertheimer and Nelson 1975; Kessler-Harris 1979; Feldberg 1980; Milkman 1980). The advocates of indirect effects have not fully explained why, in the absence of male organization, rival all-female unions do not develop. The experience of the CIO shows that the absence of established union sanction does not stifle worker militancy when other pre-conditions of organization are met. However, labor force segmentation still remains an important theoretical possibility in explaining female docility.

Spurious effects are those created solely by the occupational or industrial distribution of women. Female non-participation is a surrogate for the behavior of people in weak labor market positions. Such arguments are usually difficult to disconfirm. Since men and women rarely hold jobs with identical skills or prospects, it is hard to parametricize these completely and examine pure residuals. This notwithstanding, occupational effects appear to be quite potent. Joseph Antos et al. found in an analysis of current population survey data that 57 percent of the gap between men and women in union membership could be explained by a simple decomposition into crude occupational and industrial categories using only eight industries and seven occupations. How much more would have been explained by a full decomposition into more sophisticated occupational categories can only be imagined (Antos et al. 1980). George Bain has shown that among British white-collar workers sex differences in union membership disappear with the simple introduction of firm size (Bain 1970).

Determining the causes of low female union activity is especially important to the study of white-collar labor. The relative docility of white-collar workers has sometimes been explained by the high percentage of women in these occupations (Blum 1971). Furthermore, increasing female labor force participation has been offered as a partial explanation of the decline of union membership in the United States as a whole (Barkin 1961). Finding low direct and indirect effects would weaken the case for both of these propositions.

This chapter considers the effect of sex on the activity of white-collar unions of the General Post Office and the Great Western Railway. There are two measures of worker militancy that can be considered, union membership and participation in strikes. The former measures the willingness of workers to make long-term commitments to their union, although the costs of such involvement are often low. Strike participation is more short term but represents a more explicit demonstration of opposition to an employer. Both are valid, complementary measures of militancy; in general, studies of white-collar organization have emphasized union membership rather than strike participation, due to the relatively small number of white-collar strikes that can be analyzed.

The two firms in this study allow a test of the effect of sex on both measures of union activity. Union membership data exist both for the civil service and for the clerical force of railways as a whole. Detailed strike rates are available for the two most important strikes of the period, the 1871 postal telegraph strike and the participation of the Great Western Railway Railway Clerks Association in the General Strike of 1926. These two actions were probably the largest and most important white-collar strikes in Britain before 1945.

The reader who is primarily interested in contemporary occupational sex-typing may question whether strike rates from the early twentieth century are at all germane to a test of women’s role in reducing labor militancy today. Actually, the early date of these strikes is a very attractive methodological asset. Strikes before World War II provide a strong test of the direct effects of sex in inhibiting participation in unions. In the Victorian era and the early twentieth century, traditional sex roles were very strong. If repressive female socialization and limited labor force participation has any effect on strike activity, such an effect should be particularly manifest in early data. If no direct effects are found in these data, it is even less likely that direct effects would be found in modern data, where the impact of traditional female roles would be far less pronounced. Thus, these cases allow for a test of the effect of sex on industrial conflict that is biased towards a finding of strong non-spurious effects.

The analysis suggests that after occupational barriers to organization are taken into account, the differences in militancy between men and women were not very pronounced. In the case of union membership, there were simply no differences between men and women. David Lockwood has analyzed the union membership data in his book The Blackcoated Worker. He found that, in both the RCA and the civil service unions, the ratio of female to male membership in the unions was exactly proportional to the ratio of females to males in the relevant occupations (Lockwood 1958). Lockwood’s analysis of these materials is entirely reasonable, and there is little reason to belabor these findings here.

For strike data, the situation is more complex, but the general findings are similar to those of Lockwood. In the Post Office, the raw data show men and women striking at similar rates. This is despite substantial sex-specific barriers to mobilization. In the GWR, there were raw differences in the strike rates of men and women. However, controlling for a few sex-neutral occupational characteristics reduces the differential to a very small amount. Furthermore, much of the remaining gap can be attributed to occupational power rather than gender per se. Direct and indirect effects are not explicitly measured here. However, the residual sex differences are so small after simple job-related characteristics are taken into account that the direct and indirect effects are likely to be miniscule.

The General Strike of 1926

The evidence about the behavior of railway clerks during the General Strike of 1926 comes from the Chief Mechanical Engineer’s Department in Swindon. For every office we know the sex composition, the job structure, and the extent to which the men and the women in the office went out on strike. Because of the extraordinary diversity of the Chief Mechanical Engineer’s Department with its combination of shop offices, secretarial offices, and large concentrations of clerical semi-proletarians, there is a rare opportunity to control for objective differences in the organization of work, to see how the labor process interacts with sex in determining rates of clerical militancy.

Of all clerks, 44.7 percent struck. Of all male clerks 50.6 percent struck, but only 26.9 percent of all female clerks struck. There was, thus, a difference of 23.7 percentage points between men and women in the likelihood of striking. This is not an extreme effect, but the sex difference is fairly substantial. The difference is statistically significant by most normal criteria. However, the sex gap is severely overestimated. Women were disproportionately concentrated in offices that suffered from non-sex-based obstacles to effective organization. When one controls for these organizational impediments, the sex differences decrease dramatically.

To understand the analysis that follows, it is helpful to elaborate the different kinds of work settings that existed in Swindon in some detail. The biggest difference was between workers who worked in the shops and workers who worked in the administrative buildings. The shop clerks were the clerks who worked on the actual premises in which factory work was done. Other clerks worked in office buildings that were purely clerical in function. The shop clerks were divided into three types: administrative, stock-checking, and factory floor. The shop administrative offices did routine clerical work and had no contact with the factory operations themselves. Workers were wholly isolated from both the production process and the blue-collar workers on the floor. The factory floor offices were just the opposite. They were desks located within the factory where clerks recorded output and circulated job orders. These clerks had intimate contact with blue-collar laborers and were integral parts of shop-floor personnel networks. In an intermediate status were the stock checkers. Their work involved record-keeping in the front offices of the factory. However, inventory control required constant trips onto the production floor to examine supplies. They thus had an intermediate amount of blue-collar contact and intimacy.

The shops were also divided into different buildings. Each building had its own level of militancy. There was a carriage and wagon works, a sawmill, and a locomotive works. The sawmill and the locomotive works were quite radical, the carriage and wagon works comparatively conservative.

There were enormous differences, as well, among the non-shop offices. Most of these were located in one central administrative building, which was located in the complex with the rest of the mills. The addressograph operations and the train statistics division were located at some distance from the main complex in an old railway depot nearer town. This isolated subunit consisted largely of unskilled machine operators. They were overwhelmingly female, with a small number of both male and female supervisors who shuttled back and forth between the town station and the main building.

Within the main building there was also a mix of office types. There were some virtually all-male offices, containing no women or having a single female secretary. These usually dealt with technical and engineering matters. There were a variety of larger offices having several work groups, each with a minority of female clerks. These dealt with a variety of administrative matters, with the women doing general clerical work. Finally, there were some large, heavily female offices resembling the town station. These offices employed large numbers of replaceable clerical proletariats and the work usually involved simple statistical calculations. Arithmetic processing may have been more mechanized at the town station; nevertheless, both types of heavily female offices specialized in the more menial aspects of quantitative data-processing (Staff Charts 1926).

Women were not randomly distributed among the offices, but were concentrated in offices that were prone to passivity. First, women were less likely to be employed in offices that allowed close contact with manual workers. Stock clerk posts and factory floor offices provided for blue-collar contact. Only one woman was employed in such a situation. The importance of blue-collar contact for railway clerical organizing can hardly be underestimated. The Railway Clerks Association largely grew as a by-product of the National Union of Railwaymen. The railwaymen’s victories lowered management’s resistance to collective bargaining for the clerical staff. Furthermore for some grades, such as some blue-collar supervisory positions, the boundary line between NUR and RCA coverage was ambiguous enough to warrant jointly coordinated and conceived bargaining offensives.

Furthermore, the general strike of 1926 was a political strike induced by developments in mining. The NUR was one of the major sponsors of the strike nationwide. The NUR had direct links of cooperation to the craft and industrial unions working in the Swindon shops. Thus the prime momentum for mobilization for the General Strike would have come from the Swindon shop floors (Bagwell 1963; Tucket 1976). Anyone working in such a setting could not have avoided being swept up in the preparations for this major offensive, and shop-floor contact would have been a major stimulant to radical activity.

Second, women were concentrated in buildings with bad organizing climates. The extreme cases were the town station, the locomotive works, and the sawmill. The town station was several miles from the main complex. Any organizing drive by the RCA among the main body of railway clerks would have missed the workers at the town station. Furthermore, the workers at the town station were very unskilled. The women were entirely machine operators. Most women in the main building would have had significantly more market power. The town station thus for both organizational and occupational reasons would have been an infertile source of union activities. Women were disproportionately concentrated in this sector.

In the locomotive works and the sawmill, on the other hand, all the clerks had either direct shop-floor contact or indirect contact through working in close proximity to visitors to the shop floor. Workers in these buildings were very militant, but very few women were employed here. The scarcity of women in these settings helped to reduce levels of overall female militancy.

The main building, where most women were concentrated, had intermediate levels of union activity. These clerks would have been cut off from the blue-collar mobilization on the shop floor but would have been in the mainstream of the RCA’s own campaigns. The carriage and wagon mill was also very conservative. The reasons for this are not easy to determine. It could have been due to lower militancy among wood-working locals than among the boilermakers and machinists. It could have been due to differing plant management policies. This was an exception to the pattern of female concentration in passive areas, being a heavily male enclave that happened to be conservative. Thus, both building and shop-floor contact had independent effects on levels of militancy. How these interacted with sex can be seen in Tables 7.1–7.3.

The tables show male, female, and total strike rates for workers in every building both with and without direct shop-floor contact. Workers in stock offices and those actually located on the shop floor are defined as having shop responsibility. Before considering the impact of sex itself, we should note how the sex-specific rates confirm the discussion of the roles of the control variables. Within each building, those workers with direct shop-floor contact are more likely than their counterparts to strike. The size of the differential varies from 2 to 28 percentage points.

Clerks with intermediate amounts of shop-floor contact had intermediate strike rates. Stock clerks spent part of their time on the shop floor. Fifty percent of the stock clerks struck. This figure is between the 58.0 percent shop-floor figure and the 45.8 percent non-floor figure.

The buildings also affected militancy. For each sex, carriage and wagon workers were the least radical, town station workers were the next most conservative, main office workers were in the middle, and locomotive and sawmill workers were the most radical. This holds for both sexes and for workers with and without direct shop-floor contact.

None of these controls is a surrogate for gender. If building or shop-floor effects were entirely due to sex composition, then one would find no building or floor effects after controlling for sex. However, these differentials exist within sex categories. This strongly suggests the independent importance of building and shop-floor contact. Since women had no shop-floor contact and were located in nonmilitant buildings, these factors would have reduced overall female strike rates.

Because of this, there were large raw differences between male and female strike rates. As we saw, 50.6 percent of all male clerks struck, but only 26.9 percent of all female clerks. The 23.7 percent difference is not overwhelming, but it is statistically significant by most criteria. However, after one takes the control variables into account, the difference becomes much smaller. The best estimate of the effect of the sex gap net of building and shop-floor contact is a 7.4 percent difference. This is one-third the size of the originally observed 23.7 percent difference. A 7.4 percent gap is what would be observed if 47.4 percent of the men, but only 40.0 percent of the women, struck. The difference is real and statistically significant, but is, in absolute terms, fairly small.

Before examining the precise derivation of the 7.4 percent estimate, we should consider the differences between male and female workers who had identical building location and shop-floor contact. Large differences between men and women are noticeable in the locomotive works, the carriage and wagon works, and the town station. Of these three contrasts, however, only two show women being more passive. In the locomotive works, in fact, the women were more strike prone than the men. Furthermore, one of the cases of relative female non-striking, the town station, is overstated. In the town station, most of the men shuttled between the town station and the main building. Most of the women were isolated in the town station. There is a 42.9 percent difference between the sexes, but that is because the men’s rate resembles that of their counterparts in the main building. Differential contact with the mainstream of organizing probably accounts for much of this difference.

All three of these contrasts are based on small numbers of either males or females. The main building is the only contrast with large numbers of both. In the main building there was a 9.9 percent difference in strike rates favoring the men. This is one-third the size of the gross differential and is objectively quite small. This estimate is close to that obtained from the combined data.

To obtain an estimate of sex effects net of control variables that uses all of the data, an indirect standardization was performed. Assume that there are no sex differences in strike rates whatsoever after building and sex are controlled. We can then simulate what the overall male and female strike rates would have been. The difference between the sex gap calculated this way and the observed gender gap is the gender gap net of building and shop-floor contact.

To perform the indirect standardization, we assume that men and women in each building-contact category would have struck at the combined sex-blind rate. These sex-blind category-specific rates are listed in Table 7.3. The two sexes differ in the percentage of each that would be subject to each rate. For example, 5.5 percent of the men but only 1.9 percent of the women were located in low-contact locomotive works offices. The indirect standardization involves multiplying every rate in Table 7.3 by the percentage of workers of each sex subject to each rate. Adding up these weighted figures for males and females separately gives an estimate of the strike rates that allows for compositional but not within-category sex effects.

By these calculations, men would have had a strike rate of 48.8 percent and women a rate of 32.5 percent. This represents a gender gap of 16.3 percent that is attributable exclusively to the different distribution of men and women among buildings and differential access to blue-collar workers. Subtracting 16.3 percent from the original gap of 23.7 percent provides the sex differences due to all other causes. This is 7.4 percent.

It should be noted that the 7.4 percent difference, while small, is statistically significant. A logistic regression containing sex and both control variables provides significant coefficients for all variables (Table 7.4). No variable can be removed without a significant loss of overall goodness of fit. This also reconfirms that the control variables act independently of sex.

However, 7.4 percent is an overestimate of the size of the direct and indirect sex effects. This is because some of this difference is due to the difference between sexes in market power. Since menial jobs carry little market power, regardless of which sex holds them, estimates of the true gender effect must be purged of as much of the effect of market power as possible. Even within the control categories already used, occupational power still predicts some unexplained variance in the strike propensity of the women clerks.

The women clerks could be divided into two groups, the secretaries and the single-task operatives. A secretary is a single woman with primary responsibility for providing a variety of clerical services to a small number of individuals. The secretary’s most important task usually involves typing reports and correspondence. However, these can expand to include dictation, telephone work, and ultimately participation in the administrative affairs of the office. Single-task operatives are the workers invoked in popular stereotypes of clerical proletariats. These are workers in large, centralized, depersonalized offices doing routinized and repetitive work. Such workers are usually given a single primary task: keypunching, addressographing, tabulating, or making ledger entries. While there is some division of labor, each worker’s tasks are essentially similar to those being done by her officemates. Large insurance and accounting offices are major employers of such labor.

Secretaries have substantially more market power than do single-task operatives. This comes from the personalized nature of the relationship between the secretary and her employer and the ambiguous definition of secretarial job duties. Secretaries are expected to handle whatever clerical and minor administrative problems their employers are likely to be faced with. Proficient secretaries, as they gain experience, usually learn to provide a variety of services beyond merely typing dictated letters. This can entail providing information on commonly used files, composing routine statements for frequently sent letters, or making practical arrangements for the solutions of petty problems. Rosabeth Moss Kanter has argued that, in some cases, this can mean the substantial transfer of responsibility from executive to secretary. Examples involve those secretaries who graduate to handling some of the duties of the executive himself and secretaries who collude in the protection of deficiency, such as covering up for an alcoholic.

Kanter argues that such expansion of responsibilities is very person-specific. It depends both on the qualifications of the secretary and on the needs of the employer. Different individuals have their own routines and require different responses from their secretaries. As a result, good working relationships are often difficult to achieve. When these occur, executives work to preserve these ties by taking their secretaries with them from job to job. A good secretary becomes difficult to replace; this gives her a viable exit threat (Kanter 1977).

The situation is different for the single-task operative. Because of the lack of variation among job duties, single-task operatives are interchangeable. This standardization also allows for the easy training of new recruits from the labor market. Because of the easy replaceability of single-task operatives, the threat of withdrawal of their labor poses only minor burdens for the employer. From a position of market power, one would thus expect more resistance to employers from secretaries than from operatives.

The measure of secretarial status used in the following analysis is the sex composition of the office. Women working in offices less than one-third female are considered secretaries. Women working in offices one-third female or more are considered single-task operatives. The measure may seem counterintuitive, but it is supported by the organizational charts of the CME. The task descriptions and spatial organization of women completely differ in the two types of offices. In the heavily female offices, there are usually several work groups. There is usually one or more clusters that are predominantly male, occasionally including one or two females. These are balanced by large clusters that are very heavily female. These clusters have job titles suggesting machine operation, such as “Addressograph” or “Tabulation of Mileage Statistics.” While there are some women in these clusters whose status is ambiguous, they are outnumbered by the large pool of women in the machine subgroup.

In male-dominated offices, a different pattern emerges. One sees several clusters of male workers, each with one or two women attached to them. The job description for each cluster is complex and technical. Most involve several different duties. The job titles involve knowledge of railway administrative procedure and are often incomprehensible to outsiders. This suggests jobs requiring firm-specific training. Because of the small number of women in each cluster, and the fact that the clusters did substantively different kinds of work, the women in these groups would have been somewhat less replaceable. This makes sex composition a rough indicator of secretarial status.

Table 7.5 shows the difference in strike rates for each sex in the two kinds of offices. Due to the low variance in sex composition of offices in the town station and mills, the analysis is confined to the main building. The table shows that 41.1 percent of the women in secretarial offices struck. This contrasts with 28.8 percent in those offices where women were single-task operatives. This is a 12.3 percent difference. Note that this difference does not apply to the men. The strike rate for men in the two types of offices are practically identical. This suggests that the explanation of the strike difference among women is not to be found in some sex-neutral aspect of the offices involved. These offices had some factor that affected only the women. Variations in female market power is consistent with these differentials.

Secretarial status also enhanced the ability of women to resist opposition from male co-workers. Table 7.6 shows the female strike rates for secretarial and operative workers in offices with and without substantial male support for the strike. An office was coded as having low male strike participation if less than 45.5 percent of the men went out, 45.5 percent being the mean male participation rate in the main building. There are two major findings in the table. The first is that male support dramatically increased female strike participation rates. In secretarial offices, male support raised the female strike rate 14.4 percent. In the operative offices, male support produced a very substantial 31.3 percent difference. These findings have two conceivable interpretations. It could be that some exogenous factor such as organization jointly increased male and female rates in these offices. Or it could be that women used men as role models and imitated their strike behavior. Probably both factors were at work.

The second finding is that secretarial status increased women’s ability to strike in the face of indifferent male support. Where male support was high, secretaries were more likely to strike than operatives, but the differences were very minor. Less than 5 percentage points separates the two groups. However, there was a 21.3 percent difference in striking when male support was absent. Over one-third of the unsupported secretaries struck, while fewer than one-sixth of the unsupported operatives did so. It would thus seem that male peer support and female market power were functional equivalents in producing female militancy.

What these findings suggest is that women’s occupational characteristics to some extent determine their propensity to strike. Weak market power correlates with conservatism and passivity. We also know from the data on non-entry-level positions that women were disproportionately confined to low-status jobs. Thus, occupational differences in bargaining strength would have explained some of the differences between the sexes in strike rates. A direct test of the effect of status on the militancy of both sexes would be desirable. It is impossible with the data currently available. However, the figures presented here suggest a sensitivity to measures of bargaining strength that were only tangentially tapped by the previous control variables. Thus, the already small 7.4 percent strike difference between the sexes would have to be further reduced should occupational status be taken into account. None of the control variables explicitly measured or implied in the discussion of bargaining power are in any way direct or indirect effects of gender. Thus, the GWR data suggest negligible direct and indirect effects of gender on strike militancy.

The Telegraphists’ Strike of 1871

The telegraphists’ strike of 1871 also shows negligible direct gender effects. What is especially remarkable about this case is that this is one in which enormous obstacles were placed in the way of female participation. The strike was over issues relevant only to the men. Most of the women had been brought in explicitly as strike breakers. Furthermore, the women were isolated from most of the physical locations where organization occurred. Under such conditions, female inactivity would have been wholly understandable. Instead, the women struck at almost the same rate as men. The women did not stay out as long as the men; however, those that returned systematically frustrated management’s attempts to replace striking workers. Thus, the women served both as strikers and as strikers’ auxiliaries.

In the 1860’s the Post Office nationalized the previously private telegraph companies. With nationalization came a long and controversial reorganization of personnel policies (Cohen 1941). In the private era, telegraphists had been paid modest salaries. These, however, had been substantially increased by bonuses and the right to collect fees for certain services. The Post Office, following civil service precedent, sought to incorporate these into a basic salary. The private companies had paid on a variety of different wage scales. The Post Office sought to standardize these. However, the most immediate grievance involved promotions. The private companies had provided regular annual promotions and increases in pay. However, between the announcement of nationalization and the actual transfer of managerial responsibility, a period of two years, the companies had frozen wages. The Post Office thus inherited a labor force that perceived itself as being owed two years’ worth of back pay raises.

Postal policy immediately inflamed the situation. Since all payment was to come in salary form, the Post Office canceled all fees and premiums. They did not, however, replace these with a comparable salary. The telegraph service argued that fixing salaries would be a complex affair due to the problems of merging the different salary scales and calculating future staffing requirements. Part of the fees were remanded in the form of salary. However, the missing salary increases and the rest of the fees were to be given only after exhaustive consideration of the matter. Until a policy could be set, all salaries were refrozen.

The Post Office then took no action for two years. During this time, two more raises that the workers would have otherwise received were lost. During this period of both absolute and relative pay loss, the Post Office divulged nothing about what future salary levels would be like and made no guarantees that back pay would be given. Not even regional postmasters were given any information that could be used to comfort the men. Worker morale dropped to very low levels. Union agitation, which had existed to a small scale in the private companies, began to increase in scope and intensity.

The government was fully aware of these developments. The telegraphists communicated with each other over night telegraph lines. After hours, supervision was intense, and telegraphists could carry on private communications over the less heavily used wires. Management had become aware of this and had instituted wiretaps on lines connecting union centers. The authorities thus had perfect knowledge about the union’s leaders, centers of militancy, and future plans. Anticipating a strike of national proportion, the government began to prepare large bodies of substitute labor (Telegraph Strike 1872).

The eastern district of the telegraph service was transferred to the Royal Engineers. This was the branch of the army in charge of military signalling. The working of this section by the army was supposed to provide a training for wartime signalling operations. However, the creation of skilled telegraphists under military discipline had obvious civilian functions as well (Telegraph Strike 1872).

The Post Office also began aggressively to recruit female telegraphists. The private companies had made very sparing use of women in a small number of urban offices. The Post Office began to train women on an enormous scale. Large all-female classes were taught in a specially constructed telegraph school. A new central station was constructed in London to coordinate the facilities once handled by separate lines. The new station was staffed with a majority of women. The authorities had great hopes of female docility. When Francis Scudamore, the director of the telegraph service, was justifying this feminization to a parliamentary committee, he explicitly cited low female labor militancy (Scudamore 1871). When the Edinburgh postmaster feared an imminent strike among his men, he wrote to London requesting the implementation of an all-female instrument room (Telegraph Strike 1872).

When the strike came, it was decisively defeated. Management precipitated it by dismissing six of the union leaders. Only a few cities went out in support. Management threatened the strikers with dismissal. Five days afterwards, the workers returned with no concessions in hand. Over two hundred workers were transferred, suspended, or dismissed. However, sex played little role in the defeat. This can be seen in Table 7.7, which shows the strike rates of both male and female offices in major telegraph centers. The unit of analysis is the staff of workers in a particular office. When offices employed both sexes, the male and female staffs are considered as separate units. Thus, the Manchester men, the Dundalk women, the London men, and the London women are considered as four equal and separate cases. To keep small units from swamping the analysis, staffs are only considered if they include at least fifteen workers of a particular sex.

Male staffs and female staffs were equally likely to strike: 23.5 percent of the male offices and 20 percent of the female offices left work at some point during the action. These percentages are virtually identical. The differences in Table 7.7 are not at all statistically significant. To be sure, most women did not support the strike. Four out of five offices continued work as usual. However, this same low level of commitment was also shown by their male counterparts. The loyalty of the women was simply the loyalty typical of the telegraph staff.

One problem in the analysis is the small number of female cases. Only one female office went out: Dublin. Since the analysis rests on the behavior of these Dublin women, some closer inspection is clearly warranted. Dublin was one of the first cities to join the strike. More men went out in Dublin than in any other city in the United Kingdom. When the signal for the strike arrived, both men and women left their posts. In the course of the first day, the postmaster negotiated separately with the men and the women. By mid-day, he was able to convince most of the women to return to their stations. It thus temporarily appeared that strategy of division by sex was likely to succeed.

This proved false when the postmaster brought in Royal Engineers to replace the rest of the striking men. On the arrival of the military, the women summarily ceased work, declaring that it was unbecoming for refined ladies to work side by side with common soldiers. The postmaster was told he could use the women or he could use the Engineers, but he could not use both. An agreement was finally reached by which the military would only work at night, after the women had gone home, and the women returned to their work. The men stayed out for four days. The women had returned early. Nevertheless, it was clear from their early walkout where their sympathies lay. Furthermore, throughout the strike they continued by their noncooperation to undercut managerial attempts to replace the union men.

There were also significant differences among the women themselves. The staff consisted of four experienced women from the private company era who were highly proficient signallers. The rest of the staff dated from the recent feminization. These girls were just a few months out of telegraph school and were essentially trainees. The trainees were the ones who went back early and created the incident with the soldiers. All four of the older women refused the postmaster’s request to return and stayed out until the end with the men. Those women veterans, who had established network links to male co-workers and developed marketable skills, behaved exactly like their male compatriots. The sex differences in willingness to stay out were thus to some degree spurious effects of experience (Telegraph Strike 1872).

It is fairly remarkable that the Dublin trainees supported the strike at all; there were several considerations that should have ensured their apathy and inactivity. First, as recent entrants into the postal service, they had suffered none of the grievances germane to the strike. They had never lost any pay increases; they had never been deprived of fees or pay supplements. The concerns of the private company workers were for the women a mere matter of history. Second, the women had been hired explicitly as future strike breakers. They were visibly expected to remain aloof from combinations. Third, the women had had little time to establish social relationships with male workers. They were newcomers breaking into on-going male social networks. Fourth, the women were isolated from the physical locations where union activity occurred. The major organized activity occurred at night, when the wires permitted clandestine communication. The women were confined to day shifts, when the supervision was more intense. Finally, the women had less market strength than did the men. Labor throughout the system was generally scarce. However, female applicants were readily available while male applicants were harder to find. The expanding telegraph system gave women some job security, but they were less secure than their male counterparts.

Postal management generally viewed the feminization policy as a failure. This can be seen from subsequent gender policies in times of labor agitation. In 1880 and 1890, the Post Office faced parallel crises of strong, well-organized telegraphists pressing for improvements in their wages. Yet the late 1870’s saw the introduction of a policy of de-feminization, which continued sporadically throughout the 1880’s and 1890’s. If management had perceived feminization as being a viable tactic of labor control, it is unlikely that they would have acquiesced to such a policy. Although women’s natural docility was used by the controller of the telegraph service as part of his justification for the hiring of women, after the 1871 strike postal management dropped this line of argument from their internal discussions of female hiring.

Conclusion

The findings here support those of other investigations that have found low or insignificant gender effects. These small effects were found using a methodology that was biased towards finding large direct effects. The period chosen was one of popular acceptance of female sex-roles emphasizing docility and obedience. Furthermore, such spurious effects as differential access to promotion remained unmeasured, while occupational market power was measured only partially. Studying the modern data and using a fuller range of control variables would have made the direct effects of gender even smaller.

Two cautions should be put forward, however. One is methodological, the other substantive. The methodological warning is that the case base for this study is necessarily small. A great deal of weight has been put on the evidence from the Dublin post office and from a minority of women in unusual positions in the Great Western Railway. Further replications on other sets of early women strikers would be desirable to generate more confidence in these findings.

Second, it should be noted that the gross effects of gender were sometimes substantial. There was a 24 percent difference in male and female strike rates in the GWR. This was largely due to occupational distribution, but it still existed. The difference between gross effects and total effects has been discussed well by Paula Voos. Voos reviewed the contemporary quantitative literature on the effect of gender on union membership, dividing these studies into two types, those that studied individual predispositions to join unions and those that studied actual membership. In general, women were less likely to join unions than men. However, they were equally likely to approve of unions, to state that they would like to join a union, or to vote for unions in National Labor Relations Board certification drives. Voos suggests that these findings are inconsistent with the hypothesis that women have a low demand for unionism due to their short-term labor force participation and are more consistent with the claim that women are differentially sorted into jobs that inhibit participation in class conflict (Voos 1983). This chapter considered men and women with relatively similar although not identical occupational distributions. It is not surprising that the gross differences in union activity were small. If, however, the analysis consisted of a diverse group of workers where men were in core occupations with solid bargaining strength and women were in marginal occupations with little skill or power, then one would expect that, due to these compositional differences, the traditional gap between men’s and women’s union activities would reappear.