This thesis investigates how second year male and female English major Saudi students at King Faisal University-Hofuf (KFU) tackled two different gap-filling vocabulary achievement tests. The first is a 16 multiple choice gap-filling (MCGF) test where the subjects have to choose answers from a list of 21 alternatives above 16 disconnected sentences, while the other test is the classical cloze text with 16 non-random gapped words. These two tests were selected as a result of a preliminary study of vocabulary test types in the Saudi context, which found that these two tests are frequently used with second year university level English major male and female students in Saudi Arabia and mainly used to test knowledge of basic meaning of words. A background questionnaire, Nation’s (2001) Productive Levels Test, think-aloud and retrospective interviews were used to elicit data from 16 subjects on the effects of three main variables: test type, lexical proficiency, and gender. The verbal protocols helped in discovering what processes subjects selected intentionally to tackle test items and solve their problems during both tests. The analysis revealed a total of 86 distinct strategies in both tests along with 17 different sources of information (SI) utilised to select an answer and 8 miscellaneous behaviours during the two tests: 62 were common strategies to both tests, 7 cloze test specific and 17 MCGF specific. The ANOVA analysis showed some differences between the two tests in the common strategies, differences mainly for the better lexical proficient subjects over the poor ones, and some differences between males and females mostly in favour of males. The lexical proficiency factor was prominent over the other two variables. Almost all the strategies used in both tests are reflections of the subjects’ lexical knowledge, which show that both tests are valid measures of vocabulary knowledge.