What a great idea! To start with you seem to have a clear research question and a neat observation sheet with a set of ranking criteria for a pre-defined number of constructs to measure some sense of the status/level of instruction of science teaching at primary level. I would think about the following once the data are collected:

1. First, you will need at least more than 20 observation sheets (with all of their questions) to enable any kind of statistical testing on the data, otherwise there are two few observations, for example, for even a basic overall chi-square for one particular question. But don't worry about this component yet because...

2. You need to think about a way to summarise the data for a results section to your report, and this involves thinking about the cross-tables of average results for each observation ranking question, or group of questions in a given category, and what sorts of graphs you would like to present. For this component of the report I would sequentially conduct the following procedures:

a. Take your set of observation sheets/questionnaires and label them 1,2,3,4, 5, ...etc until each observation sheet has an assigned number.

b. Start scoring the observation sheets one ranking question at a time and write the rank score with a circle after each question.

c. On an Excel file plot the observations 1,2, 3, 4, 5, ...etc vertically in the first column and label the column 'Observation Number' or something similar.

d. Working your way horizontally across each column in the Excel file, label each column with a word to represent what each question is about, e.g., 'student participation rank' etc.

e. For each column input the score for each observation from the observation sheets 1, 2, 3,4... etc until you have a completed Excel file cross-table of scoring per observation sheet per question.

f. Now see what you can do with that data on Excel in terms of providing summary average results per question, and then plot meaningful bar graphs, until your happy with 3 main interesting graphs and the overall statistics summary.

g. Now write up the summary in word  and copy and paste the graphs over from Excel and use these summary statistics points around which to write your results section.

I have attached two links to core texts in classroom observation studies and statistics that may assist you.

https://books.google.ie/books/about/Nonparametric_Statistics_for_the_Behavio.html?id=bq3uAAAAMAAJ&redir_esc=y

https://books.google.co.uk/books?hl=en&lr=&id=dAWWk4v1wKgC&oi=fnd&pg=PP2&dq=classroom+obervational+studies&ots=m8Hujy4knG&sig=eloLpriuOhL2k9628cMpj1WEZ7A#v=onepage&q=classroom%20obervational%20studies&f=false

What a great idea! To start with you seem to have a clear research question and a neat observation sheet with a set of ranking criteria for a pre-defined number of constructs to measure some sense of the status/level of instruction of science teaching at primary level. I would think about the following once the data are collected:

1. First, you will need at least more than 20 observation sheets (with all of their questions) to enable any kind of statistical testing on the data, otherwise there are two few observations, for example, for even a basic overall chi-square for one particular question. But don't worry about this component yet because...

2. You need to think about a way to summarise the data for a results section to your report, and this involves thinking about the cross-tables of average results for each observation ranking question, or group of questions in a given category, and what sorts of graphs you would like to present. For this component of the report I would sequentially conduct the following procedures:

a. Take your set of observation sheets/questionnaires and label them 1,2,3,4, 5, ...etc until each observation sheet has an assigned number.

b. Start scoring the observation sheets one ranking question at a time and write the rank score with a circle after each question.

c. On an Excel file plot the observations 1,2, 3, 4, 5, ...etc vertically in the first column and label the column 'Observation Number' or something similar.

d. Working your way horizontally across each column in the Excel file, label each column with a word to represent what each question is about, e.g., 'student participation rank' etc.

e. For each column input the score for each observation from the observation sheets 1, 2, 3,4... etc until you have a completed Excel file cross-table of scoring per observation sheet per question.

f. Now see what you can do with that data on Excel in terms of providing summary average results per question, and then plot meaningful bar graphs, until your happy with 3 main interesting graphs and the overall statistics summary.

g. Now write up the summary in word  and copy and paste the graphs over from Excel and use these summary statistics points around which to write your results section.

I have attached two links to core texts in classroom observation studies and statistics that may assist you.

https://books.google.ie/books/about/Nonparametric_Statistics_for_the_Behavio.html?id=bq3uAAAAMAAJ&redir_esc=y

https://books.google.co.uk/books?hl=en&lr=&id=dAWWk4v1wKgC&oi=fnd&pg=PP2&dq=classroom+obervational+studies&ots=m8Hujy4knG&sig=eloLpriuOhL2k9628cMpj1WEZ7A#v=onepage&q=classroom%20obervational%20studies&f=false

Wow! Thanks a great deal. This is quite informative and useful. I am very grateful.

Excellent and comprehensive answer Michelle.

Good tip regarding new edition of Ted Wragg's book.

The only thing I would add is to increase the sample to about 50 at least, but to express overall 'trends' as percentages, I always use more than 100 responses otherwise "per cent" makes no sense. You should be able to calculate mean / mode / median and there is a statistical add-on to EXCEL called Merlin which allows you to calculate students' T test if desired and plot standard deviations and confidence levels.

Ultimately, your starting point should be to consider one hypothesis, otherwise it's an exercise in data-mining. What would you like to find out? How can you find this out from a questionnaire? How does this inform the structure / number / type of questions?

Do not try to 'prove' something since this makes you contort the data to fit your ideas. A scientist (social or otherwise) should be as objective as possible, and confirming / rejecting hypotheses is the best way to maintain this. Be happy with whatever data you get:  "rejecting a hypothesis is not failure / confirming a hypothesis is not success".

Hola hace algunos años realizamos un estudio titulado psicología de la enseñanza de la contaduría y se plantean un instrumento tipo Likert y libro de códigos. Que posiblemente te pueda dar algunos tips para tu trabajo investigativo 

 Está libre desde hace años en la Internet consultar con mi nombre o el título de la obra o mi nombre.  Saludos cordiales Marco Daza