The AAII (American Association of Individual Investors) Investor Sentiment Survey polls retail investors weekly on their market outlook (bullish, neutral, or bearish). Historically, extreme readings are contrarian signals: when bullish sentiment exceeds 55-60%, subsequent 6-month returns have averaged below normal; when bearish sentiment exceeds 50%, subsequent returns have averaged above normal. The 2023 and 2024 data showed elevated bearish readings even during a strong bull market — a pattern historically associated with continued above-average equity performance as the 'wall of worry' provided buying support.

The CBOE Put/Call Ratio measures the ratio of put options volume to call options volume on equity indices. A high put/call ratio (above 1.0-1.2) indicates heavy put buying — investors hedging against or betting on market declines — typically occurring near market lows. A very low put/call ratio (below 0.5) indicates complacency and one-sided bullishness — a contrarian warning signal. The equity-only put/call ratio (excluding index and ETF options, which are more heavily used for hedging) is a cleaner signal of speculative retail sentiment.

CFTC Commitment of Traders (COT) reports show the aggregate positioning of commercial hedgers, large speculators, and small speculators in futures markets. Historically, commercial hedgers (who use futures to hedge actual business exposure) are better informed than large or small speculators — when commercials are heavily net long, it is considered a bullish signal; when speculators are heavily net long (and commercials are net short), it is considered bearish. This 'smart money vs. dumb money' framing is overly simplistic, but the COT data provides genuine positioning transparency unavailable for many other markets.

Fund manager surveys (Bank of America Global Fund Manager Survey, conducted monthly with ~200-300 global fund managers) reveal institutional positioning: cash levels, equity overweight/underweight relative to benchmarks, sector positioning, and specific trade recommendations. When cash levels among fund managers reach multi-year highs (implying heavy underinvestment), it is historically associated with strong subsequent equity returns as managers are eventually forced to deploy capital. Historically, BofA survey cash levels above 5% have been reliable buy signals; below 3.5% have signaled elevated risk.

Natural language processing (NLP) applied to news articles, earnings call transcripts, analyst reports, and social media platforms provides real-time sentiment signals at scale. FinBERT and similar finance-specific language models classify text as positive, negative, or neutral with high accuracy. Research shows that aggregate news sentiment predicts short-term price movements with modest but statistically significant ability — particularly around earnings announcements where transcript tone carries incremental information beyond the numerical results.

Social media sentiment (Reddit, Twitter/X, StockTwits) became dramatically more relevant after the GameStop episode demonstrated that retail investor coordination on social platforms could drive significant short-term price moves. Alternative data providers now sell aggregate social sentiment scores for individual stocks — spikes in mention volume and positive sentiment sometimes precede retail-driven price momentum. However, social media sentiment is noisier than institutional sentiment data and more vulnerable to manipulation — bot accounts and coordinated campaigns can create artificial sentiment readings that mislead algorithmic trading systems.