Quant Question

S

soxboys21

New member
Joined
Jan 4, 2008
Messages
0
Reaction score
0
Nominal = observations are classified/counted with NO particular order
Ordinal = all observations are placed into separate categories and the categories are placed in order with respect to some characteristic.

Small-, mid-, and large-cap are categories, not random.
 
I dunno, I think small-, mid-, large- are ordinal. In particular, I can put ratio scale cut-offs on those categories and once you can out ratio scale cut-offs you pretty much have at least ordinal data.

The answer to the question is surely ordinal, because if they wanted the answer to be nominal they would have put something less ambiguously nominal.
 
Obviously Nominal,

Buy - Hold - Sell

Buy - Sell - Hold

Hold - Buy - Sell

Hold - Sell - Buy

Sell - Buy - Hold

Sell - Hold - Buy

What makes one better than the other?

If it were Top, Mid, Bottom, or Small Mid Large, then maybe ordinal.

Their all the same.

Nominal all the way....
 
feipar Wrote:
-------------------------------------------------------
> Obviously Nominal,
>
> Buy - Hold - Sell
>
> Buy - Sell - Hold
>
> Hold - Buy - Sell
>
> Hold - Sell - Buy
>
> Sell - Buy - Hold
>
> Sell - Hold - Buy
>
> What makes one better than the other?
>
> If it were Top, Mid, Bottom, or Small Mid Large,
> then maybe ordinal.
>
> Their all the same.
>
> Nominal all the way....

Here's the definitions, how can ^^^^^ be nominal?!?!?

Nominal = observations are classified/counted with NO particular order.
Ordinal = all observations are placed into separate categories and the categories are placed in order with respect to some characteristic.

A buy is a specific category, it's not random, etc.



Edited 1 time(s). Last edit at Wednesday, June 11, 2008 at 10:57AM by soxboys21.
 
soxboys21 Wrote:
-------------------------------------------------------
> But the buy/hold/sell still has ORDER, it's not
> random, as being NOMINAL would suggest...


like i said in previous post, it only has order in specific situations. i would think it would have to be unambiguously ordinal to be ordinal
 
why is small cap, mid cap and large cap not interval based since

small cap is <2 Bil, Mid is 2-10 bil and large cap is >10 bil ?
 
I'll clear it up for you. It's Nominal.

NOIR order from weakest to strongest. No order inthese, just three different rec's.
 
Scales of Measurement
Data comes in various sizes and shapes and it is important to know about these so that the proper analysis can be used on the data. There are usually 4 scales of measurement that must be considered:

1. Nominal Data
* classification data, e.g. m/f
* no ordering, e.g. it makes no sense to state that M > F
* arbitrary labels, e.g., m/f, 0/1, etc

2. Ordinal Data
* ordered but differences between values are not important
* e.g., political parties on left to right spectrum given labels 0, 1, 2
* e.g., Likert scales, rank on a scale of 1..5 your degree of satisfaction
* e.g., restaurant ratings

3. Interval Data
* ordered, constant scale, but no natural zero
* differences make sense, but ratios do not (e.g., 30�-20�=20�-10�, but 20�/10� is not twice as hot!
* e.g., temperature (C,F), dates

4. Ratio Data
* ordered, constant scale, natural zero
* e.g., height, weight, age, length
 
NOMINAL DATA AND NOMINAL SCALES
As implied by the title, this type of data has been given names or labels (nominal, from the french �nom� = name). Nominal data can be counted, but no superiority or preference can be implied from the numerical value of the labels, and no arithmetic manipulations can be performed on the labels themselves.
The convention within the United States is to designate highways that lead generally in a north-south direction with odd numbers, and those that lead generally east-west with even numbers. This labeling convention is useful for quickly recognizing the general direction of a numbered highway and for easily counting the number of north-south highways going through a state by quickly counting the number of odd numbered routes. This is clearly nominal data. The data (number of north-south highways) can be counted, but no superiority is implied by the numerical designations (route 95 is not better than route 5 nor worse than 101), and no arithmetic can be performed with the labels (route 101 plus route 5 does not equal route 106). Nominal data has no scale in the conventional sense that a higher number is superior to a smaller number.
In developing a technology investment strategy to combat the supply of illegal drugs, an analyst decomposed the process into hierarchical schema. The illegal drug problem results from the production, wholesale, retail distribution, and resulting generation of capital. The production process further decomposes into growing, harvesting, and processing. Wholesaling depends on transportation (from the producing country to the United States) and entry into the United States. The act of retailing depends on distribution of the drugs to the street vendors and the actual sale to users. The capital generated can be banked, laundered, or reinvested to continue the drug cycle. This hierarchical decomposition is shown in Figure 1.
Figure 1. Hierarchical Decomposition of Drug Supply
 
the name implies �order� counts, and the size of the number indicates superiority and provides rank. Webster�s Dictionary defines ordinal as �of a specified order or rank in a series� and an ordinal number as �a number designating the place (as first, second, or third) occupied by an item in a ordered sequence.� The numerical value of ordinal data indicates its relative position or standing among the data set, however, the interval between the numbers (the scale) is arbitrary, need not be consisted, and is therefore meaningless. Consequently, ordinal data cannot be combined arithmetically.
The assignment of the numerical values of 3 = superior, 2 = good, and 1 = average, creates ordinal data. It is incorrect to assume that �good,� with a numerical value of two, is twice as important or valuable as �average� which has been assigned the numerical value one. The scale, a one unit interval between values, is arbitrary and could just as well have been, Superior = 648, Good = 50, and Average = 46. Because the interval is arbitrary, the values cannot be combined arithmetically. In the first case, two �Goods� (2+2) would exceed one �Superior� but in the second case two �Goods� (50+50) would remain well below one superior. The result of such arithmetic clearly is nonsensical.
Ordinal data is useful because it clearly shows relative rank among data points. An ordinal scale is invariant under monotone increasing transformations. The numerical values that represent ordinal data indicate relative superiority and rank but an ordinal scale does not indicate by how much one factor is preferred over another. Ordinal data cannot be usefully combined arithmetically.
 
Now that I think about the answer, I do believe it is ordinal, though I answered nominal.
Buy - undervalued
Hold - properly valued
Sell - Overvalued
In terms of equity research, these recommendations are all derived from numerical comparison i.e. Stock A is overvalued by 2% above its intrinsic value, therefore we can assign numerical qualities to these stocks, although direct comparison won't derive much meaning.
Wish I had thought of this during the exam.
 
you sure did add a lot of context to the original questions that wasn't there before to get to your conclusion that its ordinal. who said overvalued was in any order compared to the other two. it might be for a long only mutual fund. but the order could be reversed for a short fund or a long/short. its nominal until you add context
frangoya Wrote:
-------------------------------------------------------
> Now that I think about the answer, I do believe it
> is ordinal, though I answered nominal.
> Buy - undervalued
> Hold - properly valued
> Sell - Overvalued
> In terms of equity research, these recommendations
> are all derived from numerical comparison i.e.
> Stock A is overvalued by 2% above its intrinsic
> value, therefore we can assign numerical qualities
> to these stocks, although direct comparison won't
> derive much meaning.
> Wish I had thought of this during the exam.
 
besides the fact that buy doesn't have to mean "undervalued". it could be based on event predicting, momentum, etc.
ryanwtyler Wrote:
-------------------------------------------------------
> you sure did add a lot of context to the original
> questions that wasn't there before to get to your
> conclusion that its ordinal. who said overvalued
> was in any order compared to the other two. it
> might be for a long only mutual fund. but the
> order could be reversed for a short fund or a
> long/short. its nominal until you add context
> frangoya Wrote:
> --------------------------------------------------
> -----
> > Now that I think about the answer, I do believe
> it
> > is ordinal, though I answered nominal.
> > Buy - undervalued
> > Hold - properly valued
> > Sell - Overvalued
> > In terms of equity research, these
> recommendations
> > are all derived from numerical comparison i.e.
> > Stock A is overvalued by 2% above its intrinsic
> > value, therefore we can assign numerical
> qualities
> > to these stocks, although direct comparison
> won't
> > derive much meaning.
> > Wish I had thought of this during the exam.
 
I really would bet huge that the answer is ordinal. In the end, ryan has this really good point ( though I think he missed the question). The measurement characteristics of the data really only exist when you start doing something with it. The "Buy"/"Sell"/"Hold" thing are just brokerage rankings. Sometimes they even add extra categories like "Strong Buy" (is there a "Strong Sell"?). They might as well rank them 1, 2, 3 which everyone would agree is ordinal.

However, suppose that you are doing an analysis of the relationship between "Buy"/"Sell"/"Hold" and market cap. Nearly everyone would just do an ANOVA treating the BSH thing as nominal data.
 
i actually answered ordinal on the test. ha!
JoeyDVivre Wrote:
-------------------------------------------------------
> I really would bet huge that the answer is
> ordinal. In the end, ryan has this really good
> point ( though I think he missed the question).
> The measurement characteristics of the data really
> only exist when you start doing something with it.
> The "Buy"/"Sell"/"Hold" thing are just brokerage
> rankings. Sometimes they even add extra
> categories like "Strong Buy" (is there a "Strong
> Sell"?). They might as well rank them 1, 2, 3
> which everyone would agree is ordinal.
>
> However, suppose that you are doing an analysis of
> the relationship between "Buy"/"Sell"/"Hold" and
> market cap. Nearly everyone would just do an
> ANOVA treating the BSH thing as nominal data.
 
you call the truth nothing?
soxboys21 Wrote:
-------------------------------------------------------
> LOL--all of this for nothing! =)
 
"A nominal scale is a list of categories to which objects can be classified." You can classify these as buy, hold sell. It is no doubt nominal.

You can't rank those because they are not related. You can't tell me that a buy is better than a sell, what if I was short?
 
You must log in or register to reply here.
Back
Top