Last Friday, tech stocks suffered a sharp reversal following a note from Goldman Sachs (GS) arguing that these stocks have shot up too much too quickly. The Nasdaq 100 (QQQ) is up over 17% even after the downturn, but the selloff continued into this week; if it continues, those gains could disappear swiftly.

Why the concern? The large run-up indicates heavy risk-on buying by investors without a commensurate analysis of potential risks. The measure of this tendency is volatility. Goldman Sachs in its note observed that volatility in Facebook (FB), Amazon (AMZN), Apple (AAPL), Microsoft (MSFT), and Google (GOOG, GOOGL) was less than consumer staples and utilities sectors, which are traditionally known for their stalwart and slow appreciation.

There is a lot to criticize in this analysis. For one, the time frame; we’re only six months into 2017 and short-term jumps in stocks don’t necessarily indicate anything about the future, either up or down. An even easier criticism would be the “bait-and-switch” rhetoric of Goldman’s note; notice that the “FAAMG” stocks lack Netflix (NFLX), which skews the results considerably, although “FAANG” has been the historically normal metric for analyzing tech exuberance. There is also the observation that other big tech names like Nvidia (NVDA) are extremely volatile and their radical price swings indicate greater awareness of risks and market turbulence than the supposedly smooth upward trend of the FAAMG stocks.

Finally and most crucially, Goldman’s argument could be seen as a “guilt-by-association” argument in which all tech stocks are painted by the same brush, but the reality of the fundamental businesses of these companies is quite different and deserves analysis on its own. This demands a more fundamentals financial analysis to determine if price jumps are a result in incremental improvements in revenue or earnings growth; that analysis was notably absent from Goldman’s note.

On the issue of fundamental valuation of tech stocks, Citron Research has gained some recent prominence, including a New York Times profile, coinciding with the increasingly famous short-seller’s cautious report on NVDA. Among the firm’s observations are some interesting data points, such as a quarter-over-quarter increase (3%) of inventories despite an 11% quarter over quarter decline in sales. This, Citron says, is “another sign of a business chasing a stock price.”

Interesting starting point for further research. An analyst should go back to Edgar with that point and start looking into whether there’s also a year-over-year increase in inventories and annual decline in sales and if this trend is seen elsewhere in other comparable companies or if it is unique to NVDA. A look into the company’s product pipeline may also be helpful in determining why there’s a disconnect between inventories and sales (if there is).

A similar closer look into the company’s increase of accounts receivable to 46 days from around 35 days and a deceleration in growth in the company’s gaming segment would also possibly confirm or refute Citron’s call. Either way, an analyst can use the starting point of Citron’s research to begin doing a more fundamental study to make up his or her own mind. Doing this with the Goldman note can also possibly uncover which companies are being overbought, which are oversold, and whether these notes of caution are causing a much needed correction or a temporary dip worth buying.