Trecile calculation

[ALKEBSEE RADWAN]

Hello everyone
kindly I want to follow a paper regrading the analysis
the author used trecile method to assign whether the media coverage is high or low. I have the raw data but I don’t know how to calculate the trecile of the media coverage please
who can help to explain how to calculate the trecile of this variable?
It is needed

[Arthur Morris]

There are a couple ways to do this.

You could rank the variable using egen rank and then assign the terciles.

* create terciles of price
sysuse auto // example dataset
egen pricerank = rank(price) // add ', by(something )' here if you'd like to group by something
egen pricecount = count(price) // add ', by(something )' here if you'd like to group by something
generate pricepercent = pricerank/pricecount
generate tercile = 3 if pricepercent > 2/3
replace tercile = 2 if pricepercent <= 2/3
replace tercile = 1 if pricepercent <= 1/3

[Arthur Morris]

Another approach is to use the xtile function provided by egenmore:

ssc install egenmore
sysuse auto
egen tercile = xtile(price), nquantiles(3)

After you install the function you can consult the help file for more information.

[William Lisowski]

Another approach is to use the xtile command built into Stata.

xtile tercile = price, nquantiles(3)

. sysuse auto
(1978 Automobile Data)

. egen tercile1 = xtile(price), nquantiles(3)

. xtile tercile2 = price, nquantiles(3)

. assert tercile1==tercile2

.

[Nick Cox]

We're taking "trecile" to be a typo for tercile. Another term often met -- my impression is more often -- is tertile.

For a list of similar terms, with further discussion, see https://www.stata-journal.com/articl...article=st0465 pp.1070-1071 -- and also https://stats.stackexchange.com/ques...half-a-percent for an extra term "trentile".

Further contributions gratefully received.

[ALKEBSEE RADWAN]

There are a couple ways to do this.

You could rank the variable using egen rank and then assign the terciles.

* create terciles of price
sysuse auto // example dataset
egen pricerank = rank(price) // add ', by(something )' here if you'd like to group by something
egen pricecount = count(price) // add ', by(something )' here if you'd like to group by something
generate pricepercent = pricerank/pricecount
generate tercile = 3 if pricepercent > 2/3
replace tercile = 2 if pricepercent <= 2/3
replace tercile = 1 if pricepercent <= 1/3

Thank you so much, I got it, but i need to understand the meaning of
nerate pricepercent = pricerank/pricecount
generate tercile = 3 if pricepercent > 2/3
replace tercile = 2 if pricepercent <= 2/3
replace tercile = 1 if pricepercent <=1/3.
other words what the meaning of tercile 3?
because my data is not amount of money it’s just number of times ( May maximum value is 5)
please explain it to be more clear to me?
once again thanks

[ALKEBSEE RADWAN]

We're taking "trecile" to be a typo for tercile. Another term often met -- my impression is more often -- is tertile.

For a list of similar terms, with further discussion, see https://www.stata-journal.com/articl...article=st0465 pp.1070-1071 -- and also https://stats.stackexchange.com/ques...half-a-percent for an extra term "trentile".

Further contributions gratefully received.

Gratitude

[ALKEBSEE RADWAN]

Another approach is to use the xtile command built into Stata.

xtile tercile = price, nquantiles(3)

. sysuse auto
(1978 Automobile Data)

. egen tercile1 = xtile(price), nquantiles(3)

. xtile tercile2 = price, nquantiles(3)

. assert tercile1==tercile2

.

Thanks a lot
I will apply it to make sure whether it gives same results
thanks

[ALKEBSEE RADWAN]

Another approach is to use the xtile command built into Stata.

xtile tercile = price, nquantiles(3)

. sysuse auto
(1978 Automobile Data)

. egen tercile1 = xtile(price), nquantiles(3)

. xtile tercile2 = price, nquantiles(3)

. assert tercile1==tercile2

.

Thanks a lot
I will apply it to make sure whether it gives same results
thanks

[ALKEBSEE RADWAN]

Another approach is to use the xtile function provided by egenmore:

ssc install egenmore
sysuse auto
egen tercile = xtile(price), nquantiles(3)

After you install the function you can consult the help file for more information.

kindly how can i assign the top trecile ?

[Arthur Morris]

kindly how can i assign the top trecile ?

As you can see there are a lot of ways to assign quantiles. It's worth carefully reviewing your data (which you could share with us: see #12 of the faq) to ensure that the approach you choose is working as you expect. If you share your attempts to assign this variable, and where you hit a snag you'll get a better answer.

The following will create an indicator for prices that are in the third (top) tercile:

sysuse auto
xtile tercile = price, nquantiles(3)
generate top3ile = (tercile==3)

You'll need to be the judge of whether this is the correct approach in your data.

[ALKEBSEE RADWAN]

As you can see there are a lot of ways to assign quantiles. It's worth carefully reviewing your data (which you could share with us: see #12 of the faq) to ensure that the approach you choose is working as you expect. If you share your attempts to assign this variable, and where you hit a snag you'll get a better answer.

The following will create an indicator for prices that are in the third (top) tercile:

sysuse auto
xtile tercile = price, nquantiles(3)
generate top3ile = (tercile==3)

You'll need to be the judge of whether this is the correct approach in your data.

thanks a lot