What statistics test should I use?

#1
So for my university dissertation I plan to do a questionnaire to find out peoples attitudes towards garden bird health (e.g. low, neutral and high level of considerations for it), the people responding to my survey will either be regular feeders, irregular feeders or non-feeders (i.e. how much people feed birds in their gardens).
I plan to find this out using mostly likert scale questions in my survey.

However, I am wondering what statistical analysis I would use to analyse the data I would be collecting. Any help?
 

Karabiner

TS Contributor
#2
I plan to do a questionnaire to find out peoples attitudes towards garden bird health (e.g. low, neutral and high level of considerations for it),
Is this how the attitude will be measured, or will such categorization (low/neutral(high)
be carried out afterwards, and if yes, based on which information and according to which criteria?
the people responding to my survey will either be regular feeders, irregular feeders or non-feeders (i.e. how much people feed birds in their gardens)
How will you determine which group a participant belongs to?
I plan to find this out using mostly likert scale questions in my survey.
Do you mean, that you will use Likert-type items which will be summed up in order to achieve Likert scale scores,
or will you just analyse single Likert-type items?

In addition, how large will your sample size be? And which additional information will you collect
about the participats (age, gender, educational level, possessing a garden etc.)?

With kind regards

Karabiner
 
#3
Is this how the attitude will be measured, or will such categorization (low/neutral(high)
be carried out afterwards, and if yes, based on which information and according to which criteria?

How will you determine which group a participant belongs to?

Do you mean, that you will use Likert-type items which will be summed up in order to achieve Likert scale scores,
or will you just analyse single Likert-type items?

In addition, how large will your sample size be? And which additional information will you collect
about the participats (age, gender, educational level, possessing a garden etc.)?

With kind regards

Karabiner
Hi, thank you for your questions.

I haven't fully planned out the questions to ask in my questionnaire just yet (I know I should probably work this out beforehand though), but it is likely that the attitudes will be categorised afterwards based on the answers to the survey questions. For example, based off questions such as "how often do you observe bird behaviour or health in your garden" with answers such as 1) never, 2) rarely, 3) sometimes, 4) often, 5) frequently etc. and so 1 & 2 will be categorised as low considerations, 3 categorised as neutral and 4 & 5 categorised as high (if that makes sense).

I will determine which group a participant belongs in by asking general questions about their feeding habits (including, how often do you feed birds), which like the above mentioned section, will be categorised into the three different groups.

Sample size I am unsure of at the moment, hoping to have around 50 participants at least for a good amount of data. Additional information I will collect will likely include age, gender, garden type/if any garden etc.

Hopefully these answers help you to understand my question a little more.
 

Karabiner

TS Contributor
#4
For example, based off questions such as "how often do you observe bird behaviour or health in your garden" with answers such as 1) never, 2) rarely, 3) sometimes, 4) often, 5) frequently etc. and so 1 & 2 will be categorised as low considerations, 3 categorised as neutral and 4 & 5 categorised as high (if that makes sense).
That is not a measure of attitudes but of behaviour. You should carefully consider what you
want to measure. And maybe you could then adapt scales from similar studies.
And I suppose that it is rather unnecessary (and artifical) to group people into
3 groups, after you have measured their attitude (or behaviour, respectively).
If you have used decent attitude (or behaviour) measure, either ordinal scaled
or even interval scaled, you can compare the 3 groups using Kruskal-Wallis
H-test (for ordinal data), or oneway analysis of variance (for interval scaled data).

With kind regards

Karabiner