Please help identify appropriate statistical treatment

[Kadiot]

I'm writing a research paper. Working title: The relationship between raising cigarrete taxes and effect on health and government collection. The main reason for this thesis is health - address the problem of the 1M new smokers, 47 diseases and 150,000 deaths a year. Sadly, legislators don't want to take on this fight for higher taxes for obvious reason. Anyway, I'd appreciate if anyone here could advice me on statistical method appropriate in my study and guide me how to do it. I'm new in Statistics. Thank you.

 

[supadrai]

It seems like you have your conclusion already, so the question is what statistical tools will help you fit the data to your desired result. Will this be a peer-reviewed paper?

 

[Kadiot]

Hi supadrai. Thank you for your quick response.
It's just the forcast I've read somewhere -1M new smokers by 2022, not a conclusion. This will not be peer-reviewed. This is a master's thesis.

A legislator says additional excise tax will destroy the "DYING" industry.

Singson says additional excise will destroy tobacco industry

If you look at the latest production data, it doesn't look like a dying industry. Looks lime another alibi in addition to revenue collection. Health and education policies in the Philippines are often drawn without research.

 

[supadrai]

I guess when you keep using loaded words with negative connotations like "alibi" and "sadly", it is not easy to believe that you want to actually do a research paper - it seems like you want to do an opinion piece and support your arguments with research. There's nothing wrong with that, per se, but it isn't a "research paper" in any meaningful sense. But I'm probably being too pedantic, so my apologies, as I should be more helpful.

It looks like you are comparing two relationships:

1. The relationship between cigarette taxes and health.
2. The relationship between cigarette taxes and government revenue.

First you need to define and explain your terms.

What is your population? Who are you studying? What are the boundaries of your investigation? A country, a city, a province? This will probably be determined by how detailed of data you can obtain - and it may only be nationally available.

What is a cigarette tax? How is it calculated (as a percent of the value of a pack of cigarettes or a the same amount for each pack)? What was the tax on a year to year basis going back as far as you need to? This will give you your first set of numbers. Is the amount of the tax really what you need to look at (when looking to health) or is it the impact of the tax on smoking rates?

What do you mean by health and how do you measure it? I would think for your paper, you should limit "health" to diseases or problems "known" to be caused by cigarette smoking. Or maybe you can, for simplicity's sake, just pick one disease, such as lung cancer. Then how will you measure health? Disease rates? Healthcare costs? Also, assuming for the sake of argument that increases in the cost of cigarettes decrease usage (another point you need to address, which may be resolved using supply-demand curves) and that decrease in cigarette usage improves health, improved health will be a "lagging indicator" for chronic smoking-related diseases - this means that the decrease in tax will not cause an immediate improvement in health of the population.

I'm going to stop here because I have no idea whether this is helpful for you. If so, I will try to write some more. Good luck with your paper!

 

[Kadiot]

First you need to define and explain
What is your population? Who are you studying? What are the boundaries of your investigation? A country, a city, a province? This will probably be determined by how detailed of data you can obtain - and it may only be nationally available.

I was thinking of a country-wide if data is avaible through FOI request. Is the sample size dependent on my statistical test?

 

[Kadiot]

What is a cigarette tax? How is it calculated (as a percent of the value of a pack of cigarettes or a the same amount for each pack)? What was the tax on a year to year basis going back as far as you need to? This will give you your first set of numbers. Is the amount of the tax really what you need to look at (when looking to health) or is it the impact of the tax on smoking rates?

My hypothesis is that the existing law called sin tax is like throwing a bucket of water into a burning building. Looks like they need to raise tobacco tax by certain amount to prevent the forecast of 1M new smokers by 2022. I'm looking at the impact of tax on smoking. If there is indeed relationship between sin tax and lung cancer rate and relationship between sinx tax and government revenue. If there will be conclusive evidence, I will advocate a more strong sin tax to protect our children. I will submit my findings to my representative as well.

 

[Miner]

You could plot responses over time noting when specific actions such as taxes changed and whether there was a change in the response associated with that. I would look at the experience of other countries also while noting that cultural difference will have an impact also. I remember seeing cigarette packs in Singapore with graphic pictures of the facial and dental damage done by cigarettes.

 

[Kadiot]

You could plot responses over time noting when specific actions such as taxes changed and whether there was a change in the response associated with that. I would look at the experience of other countries also while noting that cultural difference will have an impact also. I remember seeing cigarette packs in Singapore with graphic pictures of the facial and dental damage done by cigarettes.

Pardon my ignorance. What type of relationship is that? I need the name of statistical method. I mean..is that a correlation, a regression or what? But I'm open to whatever works. Thanks.

 

[Miner]

There are techniques (individuals control charts) that are used in industrial statistics that are relatively unknown in other branches. The linked article by Dr. Wheeler illustrates the potential of these control charts.

 

[Kadiot]

You could plot responses over time noting when specific actions such as taxes changed and whether there was a change in the response associated with that. I would look at the experience of other countries also while noting that cultural difference will have an impact also.

I think it is called Scatter plot. Am I right?