Hi everyone!
Thank you in advance for your suggestions (I'm not an expert on STATA). I have a dataset structured like this (see below for an example). My DV is a proportion of ads that talk about health (number of ads that talk about health / total number of ads; note: 1 is super rare occurrence). The years go from 2000-2023 for several product categories for 6 countries. So basically, I have data per product category per country per year from 2000 to 2023.
I would like to:
(1) generally understand whether there is an increasing trend in the proportion of ads that talk about health over time. Should I simply run regress prop_ads year to see if there is a time effect?
(2) understand if there is a significant difference between countries in the prop_ads. Should I run a linear regression on prop_ads using as IV country and inserting fixed effects for time?
(3) understand if there are significant differences between product categories in the prop_ads. Again, should I run a linear regression on prop_ads using as IV product categories and controlling for country and year?
Thank you SO MUCH to anyone who is willing to hep. I appreciate your time.
Maria
Thank you in advance for your suggestions (I'm not an expert on STATA). I have a dataset structured like this (see below for an example). My DV is a proportion of ads that talk about health (number of ads that talk about health / total number of ads; note: 1 is super rare occurrence). The years go from 2000-2023 for several product categories for 6 countries. So basically, I have data per product category per country per year from 2000 to 2023.
| year | country | product | prop_ads |
| 2000 | USA | Beverage | 0.1 |
| 2000 | USA | Technology | 0.2 |
| 2000 | USA | Sports | 0.3 |
| 2001 | USA | Beverage | 0.5 |
| 2001 | USA | Technology | 0.24 |
| 2001 | USA | Sports | 0.36 |
| 2002 | USA | Beverage | 0.8 |
| 2002 | USA | Technology | 0 |
| 2002 | USA | Sports | 0.9 |
| 2000 | Canada | Beverage | 0.2 |
| 2000 | Canada | Technology | 0.1 |
| 2000 | Canada | Sports | 0.15 |
| 2001 | Canada | Beverage | 0.69 |
| 2001 | Canada | Technology | 0.45 |
| 2001 | Canada | Sports | 0.78 |
| 2002 | Canada | Beverage | 0.23 |
| 2002 | Canada | Technology | 0.5 |
| 2002 | Canada | Sports | 0.7 |
I would like to:
(1) generally understand whether there is an increasing trend in the proportion of ads that talk about health over time. Should I simply run regress prop_ads year to see if there is a time effect?
(2) understand if there is a significant difference between countries in the prop_ads. Should I run a linear regression on prop_ads using as IV country and inserting fixed effects for time?
(3) understand if there are significant differences between product categories in the prop_ads. Again, should I run a linear regression on prop_ads using as IV product categories and controlling for country and year?
Thank you SO MUCH to anyone who is willing to hep. I appreciate your time.
Maria
