The presence of a trade union has been recognized to have a notable impact on the determination of wages and employment. More recently economists have began to study the influence of unions on hours of work. The purpose of this thesis is to study theoretically and empirically union influence on bargaining outcomes, and particularly on hours of work. I explore these ideas using an establishment level survey data within the UK electrical engineering industry from 1979 to 1984, which covers a recessionary period. Chapter 1 of this thesis provides a brief introduction and how the thesis has been organized. Chapter 2 is a survey study of unionism and hours of work. The main difference of this survey to others is that, rather than generally exploring all the union bargaining models, the focus here is on modelling a union's influence on working hours, and providing a summary of the empirical finding in the literature. Chapter 3 studies the effect of union bargaining on working hours in different union-management bargaining frameworks. Three kinds of bargaining scenario are presented: two Right-to-Manage models and one Efficient Bargaining Model. The comparative statics show that higher union bargaining power leads to lower working hours and a higher wage rate in all the models. The employment effect is ambiguous, and depends on the bargaining framework and the value of other variables. Also a comparison of this paper with other literature has been provided. Chapter 4 is an empirical study of union's influence on hours using the survey data. Chapter 5 studies the union and firm size effects on wage rates. The firm size effect on the wage rate is found to differ in the union and non-union sectors. In the union sector, the influence of size on wages is negligible or even negative. However in the non-union sector, wages increase with firm size for manual workers. The union wage differential is larger during the recessionary phase of our sample. By decomposing the union wage differential, various factors have been found to contribute differently to it, and unemployment is one of the most significant factors.