Initial data analysis

Presently my survey is still live and at the time or of writing I have 85 respondents. I am still emailing photography clubs to contribute to the survey but I thought that the number of responses that I had so far enabled me to conduct some initial analysis of the data. The majority of the respondents have come from Facebook photography groups and photography clubs.

Initial data analysis will be conducted in SPSS looking at the quantitative data to see if any themes emerge.  The first test that I conducted was a Cronbach’s alpha test.  This enables me to look at the internal reliability of objective descriptors of the images to see how closely they are related as a group (UCLA ND). I can then consider that the objective descriptor played a part in affecting the atmospherics of the image.   To do this I select all the images that exhibit a certain objective trait such as monochrome, pictures of people, landscapes etc and run the test.  The results (a high reliability) tell me that that objective element contributed to the atmosphere of the image.  To be statistically significant the result should be greater than 0.7. (UCLA ND)

The results:

Subjects is a person 0.670

Monochrome images 0.191

Particles in the image 0.665

Movement in the image 0.613 

High contrast image 0.705

Image contains the colour gold 0.633

Image containing clouds 0.552

Low key images 0.707

Images taken inside 0.711

Image taken outside 0.800

Images taken outside (not depicting weather) 0.639

Minimalist images 0.517

Light source is visible in the image 0.670

Everything in the frame is in focus 0.696

Specular highlights 0.573

Burnt out highlights 0.714

Blues in the images 0.644

Elements in the image are obscured 0.615

 

Surprisingly only three of the objective measure came back showing that these images could be grouped.  Low key images, high contrast images and images taken inside.  There are some that came close and may become significant once I analyse the final data set  

Mean and standard deviation were calculate for the data and then the images were ranked from least atmospheric to most (EZ SPSS tutorials ND)

 

Image 6 (3.17) was rated as the least atmospheric and image 13 (7.78) as the most (Scale between 1-9).  Interestingly image 13 showed the smallest standard deviation, data points more clustered around the mean (National Library of Medicine ND). This showed that there was more agreement on the level of atmospherics of this image that any other.  There was the highest level of consensus for the top four images than the rest of the group.  There was least consensus with image 12.

  It would appear from this initial analysis that there is more consensus of what is an atmospheric image is when an image is atmospheric than when it is not.