Psych 3041 Draft

...

100 emotions are essential, useful or necessary in association with sleep deprivation. Participants were to rate using a five-point Likert scale (1 = “very slightly or not at all” to 5 = “extremely”) in the first section and a three-point classification scale (1 = “essential”, 2 = “useful”, and 3 = “unnecessary”) in the second section.

Procedure

Participants were required to read the explanatory statement before starting on the questionnaire. Participants were then required to answer the questionnaires accessible through Monash University Moodle page. The questionnaires were designed to obtain the input from the participants on their perception towards types of emotions experienced during sleep deprivation. The questionnaire took approximately 20 minutes to be completed and the data collected were compiled and analysed using SPSS software.

There were three types of analyses, which are content validation, factor analysis and internal consistency reliability, carried out on the 100 items. Content validity was to determine if there was approval from the participants on which emotions are necessary for a test of behaviour-specific emotions associated with sleep deprivation. Factor analysis was carried out to detect if there was any factor that may lead to these emotions being categorised. The internal consistency reliability was measured using reliability analysis to determine if the various test items are assessing the same characteristics.

Results

To assess the association between emotions and sleep deprivation, the collected data were analysed by conducting frequency analysis, factor analysis and reliability analysis using the SPSS software. The frequency analysis was done to represent content validation on the initial item pool of 100 emotions. To determine on which emotions are deemed “unnecessary”, frequency count on each emotion was carried out. Emotions with at least 50% approval from the participants as “unnecessary” in respect to sleep deprivation were excluded on the basis of insufficient content validity. As a result, it was found that 40 items were considered as unnecessary and only 60 emotions were retained as valid items and included in further analyses.

Next, a principal-axis factor analysis was conducted on the remaining 60 items with orthogonal rotation (varimax). The Kaiser-Meyer-Olkin measure of sampling adequacy for the analysis, KMO = .95, indicating that the sample size is adequate for factor analysis. Bartlett’s test of sphericity, χ2 = 12926.465, p < .001, demonstrated that the correlations between items were sufficiently large for principal-axis factoring. In order to obtain the eigenvalues for each component in the data, an initial analysis was carried out. 50.47% of the variance was explained by four factors with eigenvalues over two combinations. However, the scree plot showed inflections that would justify retaining only Factor 1 and Factor 2 as they are the only two factors with high eigenvalues and the remaining factors are quite small. Table 1 shows the factor loadings after rotation. The items that load on Factor 1 represent pleasant emotions and the items that load on Factor 2 represent unpleasant emotions.

Lastly, the internal consistency (reliability) of the remaining items was tested using Cronbach’s alpha. Cronbach’s alpha was carried out twice for the two factors as shown in Table 2 (Factor 1) and Table 3 (Factor 2). The ‘Corrected Item-Total Correlation’ contains the correlation between the score for each item in each factor, where all items that have correlations less than .3 must be excluded. As shown in Table 2 and Table 3, all the item-total correlations are higher than .3. Meanwhile, the ‘Cronbach’s Alpha if Item Deleted’ values shows the overall reliability of the test if that particular item was removed. There are high degree of consistency between emotions in both the factors as the Cronbach’s alpha for each item removed is higher